 | This articleneeds additional citations forverification. Please helpimprove this article byadding citations to reliable sources. Unsourced material may be challenged and removed. Find sources: "Underclocking" – news ·newspapers ·books ·scholar ·JSTOR(September 2008) (Learn how and when to remove this message) |
Underclocking, also known asdownclocking, is modifying a computer or electronic circuit's timing settings to run at a lowerclock rate than is specified. Underclocking is used to reduce a computer's power consumption, increase battery life, reduce heat emission, and it may also increase the system's stability, lifespan/reliability andcompatibility. Underclocking may be implemented by the factory, but many computers and components may be underclocked by the end user. Underclocking is the opposite ofoverclocking.
For microprocessors, the purpose is generally to decrease the need forheat dissipation devices or decrease the electricalpower consumption. This can provide increased system stability in high-heat environments, or can allow a system to run with a lower airflow (and thereforequieter) cooling fan or without one at all. For example, a Pentium 4 processor normally clocked at 3.4 GHz can be "underclocked" to 2 GHz and can then be safely run with reduced fan speeds. This invariably comes at the expense of some system performance. However, the proportional performance reduction is usually less than the proportional reduction in clock speed because performance is often limited by other bottlenecks: thehard disk,GPU,disk controller,Internet,network, etc. Underclocking refers to alterations of the timing of a synchronous circuit in order to lower a device's energy needs. Deliberate underclocking involves limiting a processor's speed, which may affect the speed of operations, but may or may not make a device noticeably less able, depending on other hardware and desired use.
Many computers and other devices allow for underclocking. Manufacturers add underclocking options for many reasons. Underclocking can help with excessive heat buildup, because lower performance will not generate as much heat inside the device. It can also lower the amount of energy needed to run the device. Laptop computers and other battery-operated devices often have underclocking settings, so that batteries can last longer without being charged.
In addition to providing underclocking features, manufacturers can choose to limit the capability of a machine in order to make it more efficient. Reduced instruction set computer (RISC) models can help makers build devices that work on less power.
Underclocking can also be performed ongraphics card processor's GPUs, usually with the aim of reducing heat output. For instance, it is possible to set a GPU to run at lowerclock rates when performing everyday tasks (e.g. internet browsing and word processing), thus allowing the card to operate at lower temperature and thus lower, quieter fan speeds. The GPU can then be overclocked for more graphically intense applications, such as games. Underclocking a GPU will reduce performance, but this decrease will probably not be noticeable except in graphically intensive applications.
Newer and fasterRAM may be underclocked to match older systems as an inexpensive way to replace rare or discontinued memory. This might also be necessary if stability problems are encountered at higher settings, especially in a PC with several memory modules of different clock speed. If a PC processor is underclocked, and the clock factor or multiplier (the ratio between the processor and the memory clock speed) is unchanged, the memory will also be underclocked.
Dynamic frequency scaling (automatic underclocking) is very common on laptop computers and has become common on desktop computers as well. In laptops, the processor is usually underclocked automatically whenever the computer is operating on batteries. Most modern notebook and desktop processors (utilizing power-saving schemes likeAMD'sCool'n'Quiet andPowerNow![1]) will underclock themselves automatically under a light processing load, when the machine BIOS and the operating system support it. Intel has also used this method on numerous processors through a feature calledSpeedStep. SpeedStep first appeared on chips like theCore 2 Duo and selective Pentium models, later becoming a standard in mid to high-end Core i3, i5, and i7 models.
Some processors underclock automatically as a defensive measure, to prevent overheating which could cause permanent damage. When such a processor reaches atemperature level deemed too high for safe operation, thethermal control circuit activates, automatically decreasing the clock andCPU core voltage until the temperature has returned to a safe level. In a properly cooled environment, this mechanism should trigger rarely (if ever).
There are several different underclocking competitions similar in format tooverclocking competitions, except the goal is to have the lowest clocked computer, as opposed to the highest.[which?][citation needed]
In general, the power consumed by a CPU with acapacitanceC, running at frequencyf and voltageV is approximately[4]
TheLinux kernel supports CPU frequency modulation. In supported processors, usingcpufreq to gain access to this feature gives the system administrator a variable level of control over the CPU'sclock rate. The kernel includes six governors by default: Conservative, Ondemand, Performance, Powersave, Userspace and Schedutil.[5] The Conservative and Ondemand governors adjust theclock rate depending on the CPU load, but each with different algorithms. The Ondemand governor jumps to maximum frequency on CPU load and decreases the frequency step by step on CPU idle, whereas the Conservative governor increases the frequency step by step on CPU load and jumps to lowest frequency on CPU idle. The Performance, Powersave and Userspace governors set theclock rate statically: Performance to the highest available, Powersave to the lowest available, and Userspace to a frequency determined and controlled by the user. The Schedutil governor estimates the load through thescheduler's Per-Entity Load Tracking (PELT) mechanism.
Underclocking can be done manually in theBIOS or withWindows applications, or dynamically using features such as Intel'sSpeedStep or AMD'sCool'n'Quiet. In Windows 7 and 10, underclocking can be set within the "advanced" settings of a power management plan.[6][7]
Earlier models of theAsus Eee PC used a 900 MHzIntelCeleron M processor underclocked to 630 MHz.
Underclocking can be performed in the EFI.
Mostsmartphones andPDAs, such as theMotorola Droid,Palm Pre, and AppleiPhone, use underclocking of a more powerful processor, rather than the full clocking of a less powerful processor, to maximize battery life. The designers for such mobile devices often discover that a slower processor gives worse battery life than a more powerful processor at a lowerclock rate.They select a processor on the basis of theperformance per watt of the processor.[8]
The performance of an underclocked machine will often be better than might be expected. Under normal desktop use, the full power of the CPU is rarely needed. Even when the system is busy, a large amount of time is usually spent waiting for data from memory, disk, or other devices. Such devices communicate with the CPU through abus which operates at a much lower bandwidth. Generally, the lower theCPU multiplier (and thus clockrate of a CPU), the closer its performance will be to that of the bus, and the less time it will spend waiting.